##// END OF EJS Templates
localrepo: iteratively derive local repository type...
localrepo: iteratively derive local repository type This commit implements the dynamic local repository type derivation that was explained in the recent commit bfeab472e3c0 "localrepo: create new function for instantiating a local repo object." Instead of a static localrepository class/type which must be customized after construction, we now dynamically construct a type by building up base classes/types to represent specific repository interfaces. Conceptually, the end state is similar to what was happening when various extensions would monkeypatch the __class__ of newly-constructed repo instances. However, the approach is inverted. Instead of making the instance then customizing it, we do the customization up front by influencing the behavior of the type then we instantiate that custom type. This approach gives us much more flexibility. For example, we can use completely separate classes for implementing different aspects of the repository. For example, we could have one class representing revlog-based file storage and another representing non-revlog based file storage. When then choose which implementation to use based on the presence of repo requirements. A concern with this approach is that it creates a lot more types and complexity and that complexity adds overhead. Yes, it is true that this approach will result in more types being created. Yes, this is more complicated than traditional "instantiate a static type." However, I believe the alternatives to supporting alternate storage backends are just as complicated. (Before I arrived at this solution, I had patches storing factory functions on local repo instances for e.g. constructing a file storage instance. We ended up having a handful of these. And this was logically identical to assigning custom methods. Since we were logically changing the type of the instance, I figured it would be better to just use specialized types instead of introducing levels of abstraction at run-time.) On the performance front, I don't believe that having N base classes has any significant performance overhead compared to just a single base class. Intuition says that Python will need to iterate the base classes to find an attribute. However, CPython caches method lookups: as long as the __class__ or MRO isn't changing, method attribute lookup should be constant time after first access. And non-method attributes are stored in __dict__, of which there is only 1 per object, so the number of base classes for __dict__ is irrelevant. Anyway, this commit splits up the monolithic completelocalrepository interface into sub-interfaces: 1 for file storage and 1 representing everything else. We've taught ``makelocalrepository()`` to call a series of factory functions which will produce types implementing specific interfaces. It then calls type() to create a new type from the built-up list of base types. This commit should be considered a start and not the end state. I suspect we'll hit a number of problems as we start to implement alternate storage backends: * Passing custom arguments to __init__ and setting custom attributes on __dict__. * Customizing the set of interfaces that are needed. e.g. the "readonly" intent could translate to not requesting an interface providing methods related to writing. * More ergonomic way for extensions to insert themselves so their callbacks aren't unconditionally called. * Wanting to modify vfs instances, other arguments passed to __init__. That being said, this code is usable in its current state and I'm convinced future commits will demonstrate the value in this approach. Differential Revision: https://phab.mercurial-scm.org/D4642

File last commit:

r38591:85da230c default
r39800:e4e88157 default
Show More
bzr.py
306 lines | 11.6 KiB | text/x-python | PythonLexer
# bzr.py - bzr support for the convert extension
#
# Copyright 2008, 2009 Marek Kubica <marek@xivilization.net> and others
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
# This module is for handling 'bzr', that was formerly known as Bazaar-NG;
# it cannot access 'bar' repositories, but they were never used very much
from __future__ import absolute_import
import os
from mercurial.i18n import _
from mercurial import (
demandimport,
error
)
from . import common
# these do not work with demandimport, blacklist
demandimport.IGNORES.update([
'bzrlib.transactions',
'bzrlib.urlutils',
'ElementPath',
])
try:
# bazaar imports
import bzrlib.bzrdir
import bzrlib.errors
import bzrlib.revision
import bzrlib.revisionspec
bzrdir = bzrlib.bzrdir
errors = bzrlib.errors
revision = bzrlib.revision
revisionspec = bzrlib.revisionspec
revisionspec.RevisionSpec
except ImportError:
pass
supportedkinds = ('file', 'symlink')
class bzr_source(common.converter_source):
"""Reads Bazaar repositories by using the Bazaar Python libraries"""
def __init__(self, ui, repotype, path, revs=None):
super(bzr_source, self).__init__(ui, repotype, path, revs=revs)
if not os.path.exists(os.path.join(path, '.bzr')):
raise common.NoRepo(_('%s does not look like a Bazaar repository')
% path)
try:
# access bzrlib stuff
bzrdir
except NameError:
raise common.NoRepo(_('Bazaar modules could not be loaded'))
path = os.path.abspath(path)
self._checkrepotype(path)
try:
self.sourcerepo = bzrdir.BzrDir.open(path).open_repository()
except errors.NoRepositoryPresent:
raise common.NoRepo(_('%s does not look like a Bazaar repository')
% path)
self._parentids = {}
self._saverev = ui.configbool('convert', 'bzr.saverev')
def _checkrepotype(self, path):
# Lightweight checkouts detection is informational but probably
# fragile at API level. It should not terminate the conversion.
try:
dir = bzrdir.BzrDir.open_containing(path)[0]
try:
tree = dir.open_workingtree(recommend_upgrade=False)
branch = tree.branch
except (errors.NoWorkingTree, errors.NotLocalUrl):
tree = None
branch = dir.open_branch()
if (tree is not None and tree.bzrdir.root_transport.base !=
branch.bzrdir.root_transport.base):
self.ui.warn(_('warning: lightweight checkouts may cause '
'conversion failures, try with a regular '
'branch instead.\n'))
except Exception:
self.ui.note(_('bzr source type could not be determined\n'))
def before(self):
"""Before the conversion begins, acquire a read lock
for all the operations that might need it. Fortunately
read locks don't block other reads or writes to the
repository, so this shouldn't have any impact on the usage of
the source repository.
The alternative would be locking on every operation that
needs locks (there are currently two: getting the file and
getting the parent map) and releasing immediately after,
but this approach can take even 40% longer."""
self.sourcerepo.lock_read()
def after(self):
self.sourcerepo.unlock()
def _bzrbranches(self):
return self.sourcerepo.find_branches(using=True)
def getheads(self):
if not self.revs:
# Set using=True to avoid nested repositories (see issue3254)
heads = sorted([b.last_revision() for b in self._bzrbranches()])
else:
revid = None
for branch in self._bzrbranches():
try:
r = revisionspec.RevisionSpec.from_string(self.revs[0])
info = r.in_history(branch)
except errors.BzrError:
pass
revid = info.rev_id
if revid is None:
raise error.Abort(_('%s is not a valid revision')
% self.revs[0])
heads = [revid]
# Empty repositories return 'null:', which cannot be retrieved
heads = [h for h in heads if h != 'null:']
return heads
def getfile(self, name, rev):
revtree = self.sourcerepo.revision_tree(rev)
fileid = revtree.path2id(name.decode(self.encoding or 'utf-8'))
kind = None
if fileid is not None:
kind = revtree.kind(fileid)
if kind not in supportedkinds:
# the file is not available anymore - was deleted
return None, None
mode = self._modecache[(name, rev)]
if kind == 'symlink':
target = revtree.get_symlink_target(fileid)
if target is None:
raise error.Abort(_('%s.%s symlink has no target')
% (name, rev))
return target, mode
else:
sio = revtree.get_file(fileid)
return sio.read(), mode
def getchanges(self, version, full):
if full:
raise error.Abort(_("convert from cvs does not support --full"))
self._modecache = {}
self._revtree = self.sourcerepo.revision_tree(version)
# get the parentids from the cache
parentids = self._parentids.pop(version)
# only diff against first parent id
prevtree = self.sourcerepo.revision_tree(parentids[0])
files, changes = self._gettreechanges(self._revtree, prevtree)
return files, changes, set()
def getcommit(self, version):
rev = self.sourcerepo.get_revision(version)
# populate parent id cache
if not rev.parent_ids:
parents = []
self._parentids[version] = (revision.NULL_REVISION,)
else:
parents = self._filterghosts(rev.parent_ids)
self._parentids[version] = parents
branch = self.recode(rev.properties.get('branch-nick', u'default'))
if branch == 'trunk':
branch = 'default'
return common.commit(parents=parents,
date='%d %d' % (rev.timestamp, -rev.timezone),
author=self.recode(rev.committer),
desc=self.recode(rev.message),
branch=branch,
rev=version,
saverev=self._saverev)
def gettags(self):
bytetags = {}
for branch in self._bzrbranches():
if not branch.supports_tags():
return {}
tagdict = branch.tags.get_tag_dict()
for name, rev in tagdict.iteritems():
bytetags[self.recode(name)] = rev
return bytetags
def getchangedfiles(self, rev, i):
self._modecache = {}
curtree = self.sourcerepo.revision_tree(rev)
if i is not None:
parentid = self._parentids[rev][i]
else:
# no parent id, get the empty revision
parentid = revision.NULL_REVISION
prevtree = self.sourcerepo.revision_tree(parentid)
changes = [e[0] for e in self._gettreechanges(curtree, prevtree)[0]]
return changes
def _gettreechanges(self, current, origin):
revid = current._revision_id
changes = []
renames = {}
seen = set()
# Fall back to the deprecated attribute for legacy installations.
try:
inventory = origin.root_inventory
except AttributeError:
inventory = origin.inventory
# Process the entries by reverse lexicographic name order to
# handle nested renames correctly, most specific first.
curchanges = sorted(current.iter_changes(origin),
key=lambda c: c[1][0] or c[1][1],
reverse=True)
for (fileid, paths, changed_content, versioned, parent, name,
kind, executable) in curchanges:
if paths[0] == u'' or paths[1] == u'':
# ignore changes to tree root
continue
# bazaar tracks directories, mercurial does not, so
# we have to rename the directory contents
if kind[1] == 'directory':
if kind[0] not in (None, 'directory'):
# Replacing 'something' with a directory, record it
# so it can be removed.
changes.append((self.recode(paths[0]), revid))
if kind[0] == 'directory' and None not in paths:
renaming = paths[0] != paths[1]
# neither an add nor an delete - a move
# rename all directory contents manually
subdir = inventory.path2id(paths[0])
# get all child-entries of the directory
for name, entry in inventory.iter_entries(subdir):
# hg does not track directory renames
if entry.kind == 'directory':
continue
frompath = self.recode(paths[0] + '/' + name)
if frompath in seen:
# Already handled by a more specific change entry
# This is important when you have:
# a => b
# a/c => a/c
# Here a/c must not be renamed into b/c
continue
seen.add(frompath)
if not renaming:
continue
topath = self.recode(paths[1] + '/' + name)
# register the files as changed
changes.append((frompath, revid))
changes.append((topath, revid))
# add to mode cache
mode = ((entry.executable and 'x')
or (entry.kind == 'symlink' and 's')
or '')
self._modecache[(topath, revid)] = mode
# register the change as move
renames[topath] = frompath
# no further changes, go to the next change
continue
# we got unicode paths, need to convert them
path, topath = paths
if path is not None:
path = self.recode(path)
if topath is not None:
topath = self.recode(topath)
seen.add(path or topath)
if topath is None:
# file deleted
changes.append((path, revid))
continue
# renamed
if path and path != topath:
renames[topath] = path
changes.append((path, revid))
# populate the mode cache
kind, executable = [e[1] for e in (kind, executable)]
mode = ((executable and 'x') or (kind == 'symlink' and 'l')
or '')
self._modecache[(topath, revid)] = mode
changes.append((topath, revid))
return changes, renames
def _filterghosts(self, ids):
"""Filters out ghost revisions which hg does not support, see
<http://bazaar-vcs.org/GhostRevision>
"""
parentmap = self.sourcerepo.get_parent_map(ids)
parents = tuple([parent for parent in ids if parent in parentmap])
return parents